Identifying Patients With Atrial Fibrillation in Administrative Data

Area of Science:

• Health Informatics
• Epidemiology
• Cardiology

Background:

• Administrative data are crucial for large-scale epidemiologic and outcomes research.
• Validating administrative data algorithms for identifying atrial fibrillation (AF) is essential.

Purpose of the Study:

• To assess the accuracy of administrative data algorithms in identifying patients with atrial fibrillation (AF).
• To evaluate the utility of these algorithms for tracking AF prevalence and care patterns.

Main Methods:

• Utilized Ontario family physician electronic medical records (EMRALD) as a reference standard.
• Developed and validated an algorithm using hospitalization, emergency room, prescription, and billing codes for AF identification.

Main Results:

• The optimal algorithm demonstrated high sensitivity (80.7%) and specificity (99.1%) for AF detection.
• The algorithm revealed an increase in AF prevalence from 1.68% to 2.36% (2000-2014) and rising anticoagulation rates.
• Novel oral anticoagulant use among AF patients increased significantly from <5% to >50% (2011-2014).

Conclusions:

• Administrative data, with a validated algorithm, can reliably identify patients with atrial fibrillation (AF).
• This approach enables effective assessment of AF disease burden trends and care patterns in extensive populations.